Significance of RT-qPCR-Based Measurable Residual Disease for Optimizing the Treatment of Pediatric Acute Myeloid Leukemia

In acute myeloid leukemia (AML) genetic characteristics are an essential component for treatment decisions and disease prognosis. The assessment of measurable residual disease (MRD) using flow cytometry is a well-established diagnostic tool for adjusting treatment during intensive chemotherapy. In addition, MRD measurement based on genetic aberrations using DNA- or RNA-based approaches promises accurate and disease-specific monitoring of therapy response. The identification of gene breakpoints of fusion genes enhances a feasible MRD monitoring with even higher accuracy than flow cytometry (Maurer-Granofszky et al., 2024). Our previous research has shown that approximately 75% of the pediatric patients have traceable genetic aberrations and can be successfully monitored (Hoffmeister et al., 2024). In this study, the prognostic relevance of MRD for monitoring treatment response in different genetic subgroups of pediatric AML is investigated.

Here, we analyzed patients aged 0.1 to 18 years from the AML-BFM-2012 study or the AML-BFM-2012 /2017 registry who were diagnosed with de novo AML in Germany between January 2014 and December 2021 and had a quantifiable MRD marker. Children with Down syndrome, treatment-related AML, and acute promyelocytic leukemia were excluded. MRD diagnostics (RT-qPCR) were performed at diagnosis, after first (1st Ind) and second induction (2nd Ind) and at the end of intensive chemotherapy (last course) under standardized conditions in the central AML-BFM reference laboratory in Essen. Patients with morphologically detectable blasts (>5%) were excluded. Given the heterogeneity of the different markers and the dynamic kinetic of MRD at different time points, the threshold values were calculated by ROC analysis for each marker and each time point. Survival rates were calculated using Kaplan-Meier curves.

A total number of 286 patients with median age of 8.6 years at diagnosis and complete data of MRD during intensive chemotherapy were analyzed. The majority of patients (49%, N=140) were stratified into the intermediate-risk group, while 28% (N= 79) belonged to the high-risk group. Morphologically, AML FAB M2 and M5 were the most common.Regarding the genetic markers, the most common group was the core-binding factor (CBF) leukemias with 41% (N=115), followed by KMT2A rearrangements (KMT2Ar) at 36% (N=104), with 38 patients with KMT2A::MLLT10 and 51 patients with KMT2A::MLLT3. While the sensitivity for CBF leukemia is in the range of 10^-4 and 10^-3, KMT2Ar with ~10^-5 and NPM1 with ~10^-6 show a higher sensitivity. The ROC analysis also shows the following parameters for the different time points: 1st ind: 3.8*10^-4 (NPM1) - 4*10^-3 (CBF), 2nd ind: 3.9*10^-6 (NPM1) - 2*10^-3 (CBF), last course: 3.4 *10^-6(NPM1), 5.5*10^-5 (KMT2Ar) and 4.8*10^-3 (CBF).

MRD positivity for the CBF leukemia did not result in significantly worse survival-rates after the 2^nd Ind (MRD_neg vs MRD_pos: 5y-OS: 97%±1.9 vs. 82±3.2%, p=0.17, 5y-EFS 84±5.4% vs. 72%±10, p=0.08). This hold true in the major CBF subgroups, CBFB::MYH11 (N=36); 5y-EFS: MRD_neg 87±8% MRD_pos 71±10%; p=0.07, RUNX1::RUNX1T1 (N=60); 5y-EFS: MRD_neg 85±6% MRD_pos 73±12%, p= 0.26. It is relevant that most patients with CBF leukemia had detectable MRD levels at the end of intensive chemotherapy (CBFB::MYH11 86%; RUNX1::RUNX1T1 83%), though this had no significant influence on EFS (MRD_neg 60±21% vs MRD_pos: 81.5±5.7%, p=0.4). Similar results were observed in the NPM1group (MRD_neg vs MRD_pos: 5y-OS: 100±0% vs. 84%±8%, p=0.28, 5y-EFS: 100±0% vs. 76±9%, p=0.47). In contrast, a relatively high proportion of KMT2Ar AML achieved MRD negativity after the second induction (80%). Still, if MRD positivity was detectable, this correlated with a significantly worse outcome (MRD_neg vs MRD_pos: 5y-OS: 78±5% vs. 48%±21%, p=0.01, 5y-EFS 72±5% vs. 50±13%, p=0.02). In particular, KMT2A::MLLT3 positivity correlated with worse outcomes (MRD_neg vs MRD_pos: 5y-OS: 86±5% vs. 50%±25%, p=0.01, 5y-EFS 79±6% vs. 25±21%, p<0.001).

MRD monitoring by RT-qPCR is a reliable and feasible tool to assess treatment response. For CBF and NPM1, no significant change in prognosis was shown. In contrast, MRD positivity in the KMT2Ar subgroup identifies patients with a less favorable prognosis, so that therapy intensification (e.g. allogeneic stem cell transplantation) is indicated. Further studies are required to generate reliable results for other AML subgroups.